My real OB speaker project

Discussion in 'Speakers' started by Serious, Mar 13, 2017.

Tags:
  1. Serious

    Serious Inquisitive Frequency Response Plot

    Friend BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    1,672
    Dislikes Received:
    0
    Trophy Points:
    113
    Location:
    Munich
    @Shinwami If you've been searching for years you probably have a much better driver than I have by now. The Beta 15 is the only woofer I've heard in this setup.

    Still, I actually quite like the cheap Beta 15. The Qts and Fs parameters are almost exactly where I want them to be. The moving mass is nicely low, which is needed for two 8 Ohm woofers in parallel to be able to match the 10 Ohm Voxativ in terms of sensitivity. Moving mass does make a significiant difference in terms of efficiency, even at Fs. The sound of the woofers actually isn't too bad a match for the Voxativs. No, they're not as clean and they sound unrefined and colored in comparison, but they have a punchy, somewhat aggressive quality to them that I like quite a bit. And they don't sound as slow as most woofers do in comparison. And best of all: They're cheap. Cheap enough to be able to afford 4 of them.

    Update:
    First up, I changed the dimensions of the speakers slightly. It's slightly smaller now. The woofer baffle is only 55cm wide, instead of 60cm and they're only 77.8cm tall instead of 80cm. I did this because the Voxativ driver was slightly too high for me and the whole thing felt slightly too wide for me. The listening axis should be at 89cm height now, which is what I wanted. The slit between the two baffles is slightly larger too now at 22.4cm, because I felt the 20cm were too small.
    I tried something interesting the past few days. Mounting the woofer backwards, so the magnet faces forwards. This makes it easily tip over so have to put either magical rock or huge cored coils and vintage drivers in the back to prevent it from falling over.

    I mounted the woofer with the squishy foam on the front baffle in an attempt to move the acoustical center forward as much as possible. As it later turned out from my measurements it only seemed to move the acoustical center by about 3cm or so. I don't really understand why it didn't move more. Overall the sound was weird like that and not in a good way. At first I thought I preferred it because this finally allowed me to get proper crossover integration, but I later realized that I could get the same integration with the woofer mounted normally.
    The woofer FR seemed to be almost identical, with the exception of (I think) an 800Hz bump and faster rolloff above 1.5kHz. Overall slightly less smooth and extended FR which was to be expected, but actually not as bad as I expected. For some reason it sounded like the bass and lower midrange had lost all definition. I actually thought I was imagining things because theoretically there shouldn't be a difference at those wavelengths. My dad immediately picked up on the difference and told me that I might aswell use a subwoofer if I kept it that way. It was really odd. Bass lacked impact and seemed to come from everywhere. Like the phase was inverted, but even worse. And yes, I did invert the woofers to keep the same acoustical polarity.

    There are really only a few reasons I can think of why this might be:
    1. The cheap steel basket hurts the sound. This should be obvious. The basket isn't very open and is quite ringy. But does this make a difference at such low frequencies? Experience tells me that anything between the driver and my ears is usually a bad thing, so it probably does make a difference.
    2. Mounting it this way the squishy foam made contact with the baffle instead of the much harder material that's on the rear. Maybe the squishy foam absorbs some energy. Seems unlikely to me, but my Voxativ drivers actually came with no material to seal the basket to the baffle. My OCD kicked in and I made a sealing gasket from some 1mm thick closed cell foam to get an airtight seal. It actually sounded worse with the foam and I threw it out again after a few minutes. It sounded too damped. The foam on the drivers is much thicker and squishier, so might be even worse, at least theoretically. The effect was also similar to screws that aren't tight enough.
    3. The squishy foam also adds some decoupling to the basket. The ringy basket actually benefits from the rigid coupling to the baffle to make it, well, less ringy, but some dynamat might still be needed. When screwed with the rear facing the baffle (mounted in front of the baffle) the ringing seems almost completely gone when tapping the basket, but I think I can still hear some of it with the measurement sweep. Here's what it sounds like when I hold the woofer in my hands and tap it with my finger and then with the back of a brush: https://www.dropbox.com/s/3dm2z4xiv5hv6w4/Beta15Basket.wav?dl=0
      With the squishy foam decoupling the basket from the baffle the ringing is very audible with the measurement sweep and, in my opinion, also audible with music.
    4. The cone shape might actually affect the dispersion a bit. I remember reading that a cone focuses the wavefront a bit more compared to a dome, even in the bass region, but this really doesn't seem that likely to me. The wavelengths are absolutely huge compared to the driver size. The acoustic center also moves relative to the baffle, which could also slightly change things. But does this all matter when I'm listening 15° off-axis anyway?

    I realized when measuring that the widebander doesn't have to be as far back as I expected to get proper crossover integration. It seems like the front of the Voxativ baffle has to be 10.5cm behind the front of the woofer baffle with the woofer mounted normally or 7.5cm with the woofer mounted backwards. I initially wanted to mount the woofer backwards to be able to move the Voxativ driver closer to the slit opening because the slit seemed to affect the soundstaging noticeably when the driver was 15cm back. It's noticeably more restricted vertically, which isn't so much a bad thing, until at a certain point where it gets too much.

    The crossover is around 360Hz, which with 60cm distance between the two woofers and a first order crossover just makes it a real D'Appolito.

    Here are CSDs of the two configurations, with the delay adjusted for the 1m measurement distance (driver is about 2cm further back than for 2m distance). I couldn't get proper results at 2m distance. Note that I am not a measurement person and I don't really care about the CSDs in the upper midrange and treble region, but if you want super clean CSDs in the upper frequencies, I can provide you with that. This was also with the microphone to the left of the left driver, which gives worse results than to the right of the left driver, which is where I sit. The mid-treble peak and upper treble rolloff are essentially gone on the other side. I also measured slightly closer to on-axis than I sit. Where I sit the treble smoothly slopes downard from about 8kHz to 20kHz by about 5db.
    The region below 1kHz is what interesting in these measurements. 7.5ms Blackman-Harris 4 terms window:
    Woofer backwards:
    Woofer backwards.jpg

    Woofer normal:
    Woofer normal.jpg

    I know the CSDs actually look better with the woofer backwards, but this is mainly because the woofer has more ouput in the 500Hz to 1kHz region this way. I don't know where the 900Hz thing came from in the 2nd measurement, but this was also there with the woofer disconnected, so at least the woofers don't fuck up the midrange. With the woofer backwards the woofers seemed to add more between 500Hz and 1kHz, which is generally a bad thing. Also because of the 2kHz peak (or rather because of the steep drop after 2kHz causing a large phase shift) the woofer actually out-phases the Voxativ driver a tiny bit at 2kHz the way I have it set up right now. I got a 2nd set of CSDs that looked cleaner than these at 2kHz, but they looked worse between 1-2kHz, so I'm presenting these. Generally there should be about 1db less 2kHz than without the woofers connected. I can get rid of this by moving the driver back a centimeter or two, but I think I want to leave it like that for now. The midrange tuning is very close to how I want it. In these measurements you can see a slight 1db bump at the crossover frequency. With the woofer backwards it seemed to be exactly at 360Hz.

    Compared to how @Muse Wanderer heard them they should sound cleaner with more precise imaging and with more warmth to the lower midrange. In that case the woofer rolloff wasn't quite steep enough, but it also set in too early. Now it should be close to a real 1st order slope, which means the woofer is around 10db down at 2kHz
     
    Last edited: Jun 19, 2017
    Priidik likes this.
  2. Serious

    Serious Inquisitive Frequency Response Plot

    Friend BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    1,672
    Dislikes Received:
    0
    Trophy Points:
    113
    Location:
    Munich
    Mundorf MCoil CFC vs Jantzen Wax Coil
    Coils1.jpg

    Progress has been pretty slow, but I finally ordered both the TOTL Mundorf and Jantzen copper coils to compare them. I used 4 coils because I feel using a coil for each driver terminal sounds better (some people call this a "symmetrical" crossover). When I compared it directly it definitely sounded better to have one coil on each leg of the driver than using both on the - leg with the Ragnarok. More natural transients and better imaging. More heft. Measurably there's no difference. I think Audio Physic and Gauder build their crossovers like this. I chose the 12AWG coils because they seemed to be a good compromise between price and DCR (FR). 0.3 Ohm DCR seems to be just about enough bass for my tastes and I'm still not using super thick cables for my bass drivers. FWIW I put both coils on bubble wrap and used clothespins to connect the cables to the coils so I didn't have to solder. Both coils seemed to benefit from letting them burn in for a few days, the Mundorf probably more so than the Jantzen.

    The difference between the foil coils and the crappy cored coil I was using before was pretty big, although all three had very similar values. The changes are similar to those you get with cables, but much bigger. Both coils simply sounded much clearer and simply more transparent. It's a big effect and once you've heard it you can't go back. The differences between the two foil coils were much smaller and with different gear and speaker drivers you could lean one way or another.
    Still, when I compared them I always ended up preferring the sound of the Jantzen wax coils. It's hard to put into words exactly why, I generally found the wax coils more engaging. The German coils sounded less lively and dynamically flatter. The Jantzen coils have a bit of a fuller, more "earthy" sound. Some people might consider this a coloration, but even with paper cones I preferred it over the more ethereal sound of the CFC. The Danish coils also seemed to have a broader palette, so I do think they impart less of their own coloration. Especially bass drum hits sounded clearer with a better defined impact that was easier to locate. More thump, especially down low where the Mundorf coil seemed a bit limp at times. At times I also felt the Jantzen had a blacker background which also comes down to its more effortless dynamic ability.

    Next up is replacing the parallel resistor and cap and making the woofer baffle and U-frame more rigid.
     
    Muse Wanderer likes this.
  3. zerodeefex

    zerodeefex Grumpiest admin

    Staff Member Friend BWC
    Joined:
    Sep 25, 2015
    Likes Received:
    9,876
    Dislikes Received:
    43
    Trophy Points:
    113
    No one has told you this yet:

    The Alphas work better in practice than the Betas I've found.

    Edit: I just read this and I meant the betas are better than the alphas!

    The PAP variant of the Alphas are fucking amazing. Get those woofers.
     
    Last edited: Sep 2, 2019
  4. Serious

    Serious Inquisitive Frequency Response Plot

    Friend BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    1,672
    Dislikes Received:
    0
    Trophy Points:
    113
    Location:
    Munich
    Sure, I know I should try them. I don't like them on paper (Qts, Fs, FR, Mms), but I know you can't judge the sound by the specs. I kind of wished they sold the OB-A15SE with the better Deltalite basket and a stronger ferrite motor. I have a bit of an irrational fear of neodymium magnets and the 1.2kHz peak of the Neo variant will most likely be a problem with my crossover.

    You compared the Beta and Alpha 15 directly? Why did you prefer the Alpha? Tbh I didn't like what I heard from the woofers with the PAP 15s all that much when I heard them very briefly at High End, but the implementation and voicing is totally different.
     
  5. Priidik

    Priidik Friend

    Friend BWC
    Joined:
    Sep 27, 2015
    Likes Received:
    1,420
    Dislikes Received:
    2
    Trophy Points:
    93
    Location:
    Estonia
    What was the cored coil you used?
    The air-cored coils are usually much higher DCR raising Qts of the driver effectively.
    I use the Jantzen wax coil for midrange btw.
     
    Last edited: Mar 5, 2018
  6. Serious

    Serious Inquisitive Frequency Response Plot

    Friend BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    1,672
    Dislikes Received:
    0
    Trophy Points:
    113
    Location:
    Munich
    It was just some coil I salvaged from an old speaker I had. That one was about 1.1mH so I used a cheap 0.26mH air core wire coil I had. In sum they had very similar DCR to what I'm using now, maybe a tiny bit lower.

    I do need a low DCR to get a high enough sensitivity. High DCR lowers the Q of the filter and I also don't need a higher Qts. I've thought about using the Mundorf CFS14 coils, but since I don't need a high inductivity I can get away with air core coils.

    What caps would you use? I need 200uF in parallel to the driver so it can get expensive very quickly. For resistors I thought I'd build my own graphite resistors.
     
  7. Muse Wanderer

    Muse Wanderer Friend

    Friend BWC
    Joined:
    Oct 7, 2015
    Likes Received:
    1,102
    Dislikes Received:
    0
    Trophy Points:
    93
    Location:
    Malta
    The bass was impeccably defined on the OB-A15SE from the start on my PAPs. It was my first "holy shit" moment after I connected them to Yggdrasil-Rag. Where did that bass come from! In your setup with sloped back wall they will probably extend even deeper. The neodymium magnets may well be the secret sauce as the woofers need all the help they can get to catch up with the superfast Voxativs. That said, I did not directly compare the alphas to the betas. I can only recall being blown away by your OBs sound.

    You can always come over for a listen any time you want @Serious .... It is just a 3hr flight :).
     
  8. Priidik

    Priidik Friend

    Friend BWC
    Joined:
    Sep 27, 2015
    Likes Received:
    1,420
    Dislikes Received:
    2
    Trophy Points:
    93
    Location:
    Estonia
    Doesn't make much sense to waste expensive caps to low voltage, low frequency parallel operation.
    I'd take any polypropylene type that's cheap, only because you have relatively high crossover point, otherwise I'd use electrolytic even.
    Imo even the cheapest Jantzen film cap is passable as series cap for midrange, more than good enough for woofer bypass operation.

    I'd bet the Voxatives would love low Qts woofers, crisp uncoloured high bass as sacrifice for sub bass action.
     
    Last edited: Mar 6, 2018
  9. Poleepkwa

    Poleepkwa Friend

    Friend BWC
    Joined:
    Oct 1, 2015
    Likes Received:
    922
    Dislikes Received:
    3
    Trophy Points:
    93
    Location:
    Finland
    @Serious, do you mind sharing your crossover build? I am interested in how you put it all together.
    My current DIY speakers uses PCB, but I have read (Troels) prefers direct PtP.
     
  10. Telstar

    Telstar Bottom 1% of posters on SBAF

    Joined:
    Mar 28, 2018
    Likes Received:
    8
    Dislikes Received:
    12
    Trophy Points:
    8
    Location:
    Italy
    What’s the xmax of the voxativ? I have never heard them but I am a bit skeptical about these very expensive wideband (lowter cough cough). I cannot stand the raising FR, suitable only for horn load.
     
  11. Serious

    Serious Inquisitive Frequency Response Plot

    Friend BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    1,672
    Dislikes Received:
    0
    Trophy Points:
    113
    Location:
    Munich
    Crossover looks like this atm:
    [​IMG]

    @Telstar Xmax is supposedly 10mm, but that's most likely peak-to-peak like the AERs. With two speakers it should be enough for 100db at 40Hz or 87db at 20Hz not considering port/BLH gain or room gain. That's still not super loud and one 8" driver is definitely pretty limited for bass.

    As far as the rising FR goes, big baffles and listening off axis helps, but even in an infinite baffle without woofer support it'll still be pretty forward (like the Ampeggio Due). In my case the woofers support the widebanders throughout the midrange for some extra warmth. Even then if you prefer a laid-back sound it's not for you.
    Obviously horn loading also works.
     
    Jackork likes this.
  12. Cspirou

    Cspirou They call me Sparky

    Friend
    Joined:
    Sep 27, 2015
    Likes Received:
    4,861
    Dislikes Received:
    5
    Trophy Points:
    113
    Location:
    Northwest France
    I wouldn't say he prefers point-to-point over PCB. More that making PCBs for one off speakers is a hassle and expensive.

    Of course there's other advantages to point-to-point. The higher currents in a speaker crossover mean thicker wire, which is limiting with a PCB.
     
    Poleepkwa likes this.
  13. Telstar

    Telstar Bottom 1% of posters on SBAF

    Joined:
    Mar 28, 2018
    Likes Received:
    8
    Dislikes Received:
    12
    Trophy Points:
    8
    Location:
    Italy
    It's very good even if its' already +-, you won't be limited in the midbass like I was with the last FR I tried.
    The rising FR means that the driver was made for horn use. You can tweak it with passive or active EQ if you wish. That's the fun of making our own speakers, isn't it? :)
     
  14. sphinxvc

    sphinxvc Gear Master (retired)

    Staff Member Friend BWC
    Joined:
    Sep 28, 2015
    Likes Received:
    2,709
    Dislikes Received:
    5
    Trophy Points:
    113
    Back when you did this, was it pink noise or a specific freq?

    Voxativ state 94db @ 60hz and 100db @ 3khz for 1w/1m.

    1w @ a 10.3 ohm nominal impedance seems to require more from the amp than the typical 2.83v, when I calculated it was 3.22v*. That sound right?

    *(3.22^2)/10.3

    With drivers that don't have a typical 8 ohm nom. impedance, does one choose to abide by 1w, or 2.83v?

    I suppose the sensitivity would be very different with a 2.83v output.
     
    Last edited: Sep 2, 2019
  15. Serious

    Serious Inquisitive Frequency Response Plot

    Friend BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    1,672
    Dislikes Received:
    0
    Trophy Points:
    113
    Location:
    Munich
    I didn't actually measure with a 2.83Vrms output... that'd require hearing protection :p. Overall it's just a guess based on the difference vs the other drivers I had from FR measurements. See for example the FR on the first page vs the Fostex FE83En.

    At 11 Ohms (which is about the resistance of the driver in the midrange) it should be about -1.5db vs 8 Ohm. I think I'd measure around 97-98db at 2.83Vrms at 1m from between 500Hz to 3kHz up to 30 degrees off-axis or so. Depends a lot on the baffle size and which frequency you chose as a reference. In the upper midrange and treble (even at 20 deg off-axis) it's likely more than 100db/W.
    At 3kHz I measured around 14.5 Ohms, so 97db at 2.83Vrms is already 100db/W. The AC-1.6 definitely achieves that (and I don't measure a peak at 3kHz). Heck, even at 20kHz (where I get about a 25 Ohm) 95db/2.83V is 100db/W.

    Usually it makes the most sense to specify the sensitivity I guess, but the efficiency can be interesting to look at and compare vs other drivers.
     
    sphinxvc likes this.
  16. sphinxvc

    sphinxvc Gear Master (retired)

    Staff Member Friend BWC
    Joined:
    Sep 28, 2015
    Likes Received:
    2,709
    Dislikes Received:
    5
    Trophy Points:
    113
    Ahh, you DID mention this. I see it now + the plot. Thanks.
     
  17. Serious

    Serious Inquisitive Frequency Response Plot

    Friend BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    1,672
    Dislikes Received:
    0
    Trophy Points:
    113
    Location:
    Munich
    During my measurements for the PAP Trio 15 clones I did another measurement of just the Voxativ driver resting on about 30cm of sheep's wool, at ca 0.5 - 1m distance. For some reason that I can't quite figure out it always seems to measure way better than in a baffle this way, yet damping the baffle or damping the back wave never makes it look as good as without the baffle. It's probably related to the baffle, but I've given up on trying to make the measurements look better at the moment. Averaging all the angles in the baffle still gives a smooth FR.

    @murphythecat This is essentially what you asked for in May 2017. I already had similar measurements back then, but never mentioned it because I thought they were too good looking for what they are. Turns out maybe they weren't.

    Voxativ AC-1.6 no baffle FR:
    Voxativ AC-1.6 no baffle FR 0-10-20.jpg
    At about 30-40 degrees off-axis a suckout develops at around 3kHz which goes away again at 50-60 degrees off-axis. In that regard I think it's similar to other drivers like this (e.g. Lowthers) and it's probably because the main cone and the whizzer cone cancel eachother at those angles. There's also some excess group delay around 4kHz, between 50 and 150 μs depending on the angle.

    Here's a CSD at what should be 20-25 degrees off-axis:
    Voxativ AC-1.6 no baffle CSD 1.75ms.jpg
    On-axis there's more treble and slightly less hash in the treble. You can also see ringing at 2kHz, but overall this is a fantastic result, to the point where I still think it might be too good to be true.
    Also attached a simulation in Edge of what the FR from not having a baffle might look like. The peaks in the treble probably won't be as bad as simulated, though.

    Next up: either cap or woofer comparison.
     

    Attached Files:

    Btdk34, Lyander and sphinxvc like this.
  18. Serious

    Serious Inquisitive Frequency Response Plot

    Friend BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    1,672
    Dislikes Received:
    0
    Trophy Points:
    113
    Location:
    Munich
    Cap upgrade:
    CDE caps.jpg

    @Priidik I decided to go for CDE SF series caps with a CDE 940C as a bypass cap to replace the ECap raw I used for prototyping. Total cost including shipping around 115€. After reading this I just had to try them: http://www.humblehomemadehifi.com/Cap.html
    This is something I should've done a long time ago, but my PAP Trio clone project is what ultimately got me to buy them. I still want to replace the resistors and then I'll put the XO on a proper board. After that I'll work on bracing and damping the back of the baffle and getting better cables. Then I should finally be done.

    I was a bit suprised that even adding the bypass caps to the cheapo Mundorf caps made quite a bit of a difference. It sounded somewhat tighter. Transients seemed more sudden, less damped.
    Replacing the electrolytics with the big CDE caps was an even bigger difference, much bigger than I anticipated. They sound much more dynamic, a lot smoother, yet more controlled and cleaner. For some reason even the imaging seems to be more three-dimensional. What's interesting is that bassy music now seems to sound bassier and music that's voiced a bit lean seems to sound a bit leaner. I think that means they impart less of their own sound. Still, overall they seem to sound a bit warmer, which was a welcome change. I thought that maybe the 8AWG coils with half the DCR would've been a better choice before (that'd get me 0.5db more bass), but now I don't think so anymore.
    Adding the bypass caps to the SF caps again made them sound a bit tighter and less damped.
     
  19. Serious

    Serious Inquisitive Frequency Response Plot

    Friend BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    1,672
    Dislikes Received:
    0
    Trophy Points:
    113
    Location:
    Munich
    Outdoor PAP Trio 15 measurements

    Alright, this is not going to be super easy, but as I wrote in the PAP thread I took measurements of the PAP Trio 15 clones outdoors. I only took measurements in the MTM config with the crossover configured as the stock Leonidas series crossover so far. I think that this should be the best measuring config here, but that's just my guess.
    I measured it with the mic at 90cm height and 2m distance from the speaker baffle. On the floor I had a 25cm high and 80cm wide piece of sheep's wool (8kg total), to absorb the floor bounce. You have to put OBs on the floor since the floor acts as an important part of the baffle, but you also have to absorb the floor reflection to get an accurate response in the midrange.
    I attached an image that shows the response with and without the wool. In this case (since the woofer is so close to the wool) you lose some energy centered around 200Hz. The wool wasn't blocking the woofer directly, but it did block its reflection.

    As I usually do I tried to reference the graphs to 2.83Vrms/1m and I measured at a level 23db lower than that. Some of the measurements are noisy from the wind, which you can see in the bass region.

    This is probably the most important measurement: The FR for angles that are close to what you'd actually listen at:
    PAP Trio 15 AC-1.6 MTM - listening window.jpg
    In this case it's not too bad, but there's a suckout at 1kHz and a broad bump between 1kHz and 2kHz. This suckout gets deeper and moves down in frequency when we move closer to the Voxativ driver:

    PAP Trio 15 AC-1.6 MTM - closer to Voxativ axis.jpg
    For the measurement where I put the mic at the same height as the Voxativ driver there's a strong suckout at 500Hz. This is because the woofers are delayed by quite a bit compared to the Voxativ driver, so the two drivers cancel. I also attached the step response. The woofer looks to be delayed by roughly 0.4ms or 13.7cm.

    That also means that when we move higher up vs the speaker (i.e. standing up vs sitting) the suckout will be less severe. And indeed we see that measured at 2m height the suckout is gone and is replaced by a broad bump in the crossover region. This is equivalent to about a 30 degree angle. Obviously this is also 30° (or slightly more) for the Voxativ driver, so the treble response will be different.
    PAP Trio 15 AC-1.6 - 30° vertical vs listening angle.jpg

    So what does the woofer response look like driven from the crossover? This is actually not that easy to measure for a series crossover since you can't just disconnect one of the drivers and measure the other like for a regular parallel crossover. So I put a large piece of absorbent foam on top of the Voxativ driver, ca 10cm thick which seems to have worked quite well. This is what the woofer response looks like driven from the crossover:
    PAP Trio 15 AC-1.6 MTM - woofer response from Leonidas XO.jpg
    It's a bit concerning to me how little rolloff there is for the woofer. It's almost at full level at 2kHz. Well, it is what it is. I'd prefer a steeper crossover considering the time-delay and the resulting comb-filtering effects between the drivers.

    What's interesting is that towards the rear the response in the crossover region is actually quite a bit smoother. Here's on-axis for the front and rear axes:
    PAP Trio 15 AC-1.6 MTM - on-axis front vs rear.jpg
    Note the sharp dip at 2.2kHz and the peak at 3kHz. This is from the Voxativ driver and makes the rear response sound awful. This is why I prefer to absorb some of the rear radiated energy for reverberant rooms.

    Finally, I also attached the CSDs of the PAP Trios and of my OB. I slightly tweaked my crossover since then, but I managed to get very similar looking CSDs in the treble range, so they can be compared directly. In the CSDs you can see the suckout that appears because the drivers cancel and how the woofers are delayed relative to the Voxativ driver. I had to use a very short 1.4ms window because I got a reflection. These were taken in-room with the mic near my couch. The difference in the treble range decay is entirely from a slightly different angle, but I do think the Leonidas XO rolls off the upper treble.
     

    Attached Files:

    Last edited: Nov 6, 2019
    Lyander, crazychile and Btdk34 like this.
  20. Serious

    Serious Inquisitive Frequency Response Plot

    Friend BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    1,672
    Dislikes Received:
    0
    Trophy Points:
    113
    Location:
    Munich
    DIY Graphite resistors
    Graphite resistors.jpg

    First up: I don't recommend this to anyone. Making these 8 resistors ended up taking me about 14 hours spread over two days. You do learn to do it over time. The first 3 took me 8 hours, the remaining 5 only took about 6 hours. Working a minimum wage job for the same amount of time I could've afforded nicer resistors, not even including the 20€ I paid for graphite leads and other failed experiments. The hard part is getting them accurate, making the resistors themselves isn't too hard. I got all 8 within 1% and the total for left and right to within 0.02%, but that should always be possible with 1% resistors and it's well within measurement error anyway.
    I mostly followed Troels' tutorial here: http://www.troelsgravesen.dk/graphite.htm
    The main exception is that I used 3.15mm 6B leads and 1.5mm^2 wire. Those aren't cheap and break quite easily, so it's not the easiest way to do it. I wanted to use 8 resistors to get four times the surface area for cooling. I tested one and at 15W it already had only 60% of its original DC resistance, but it didn't yet break. You can pretty much make a lamp out of them as long as they are heated slowly. If you run too much power through them (say 40W+) too quickly they'll just explode.

    How do they sound?
    In parallel with the woofer there's not much of a difference. I thought it sounded slightly more defined and cohesive compared to the metal film and cement reistors (in parallel) I was using before.

    The interesting bit was putting them in series with the Voxativ. Here's what I wrote down for that:

    Graphite in series with Voxativ +, 0.67 Ohm
    • Persistent tizz in the treble
    • possibly slightly less dynamic and less depth to the staging
    • slight hash/grain
    • dynamics are slightly more forced/less free flowing
    Two wires soldered together
    • Slightly tizzier indeed
    • Also slightly flatter dynamically and less plankton?
    Old resistors in series with the Voxativ +, 0.5 Ohm
    • definitely less dynamic now
    • less resolving and also less stage depth now
    • treble seems smoother than graphite resistor, but wires aren't the same for both resistors
    • treble is still very tizzy compared to stock
    • background blackness lacking, even compared to graphite
    • treble has more hash, less tizz
    • upper midrange seems more laid-back compared to graphite
    I think both of the resistors sounded a lot worse than just using a piece of wire. It seems like even resistors hurt the transparency quite a bit.
     
    Last edited: Nov 13, 2019
    GanGreinke, numbercube and Priidik like this.

Share This Page